eval_toolkit.provenance#
|
Sentinel-style hit for a successful file hash. |
|
Sentinel-style miss for a file hash that could not be computed. |
|
Current HEAD full SHA, or |
|
SHA-256 hex digest of an existing file (sentinel-typed). |
|
Build a JSON-serializable provenance dict for |
|
SHA-256 hex digest of an existing file. |
|
Create and return a timestamped run directory. |